Simulator: separate static calibration and drifting-bias profiles

[tribbloid]

Source: Packages/ARGlassesSDK/__module~/driver/src/sim/TODO.md

The default simulator continuously changes hard-iron offset, so it is not a stationary ellipsoid suitable for validating a static calibration solve.

Acceptance criteria:

• Make the baseline DummyConfig::default() use zero hard-iron drift.
• Retain the existing drift amplitudes and five-minute period in an explicitly selected public adaptive-calibration stress profile.
• Document the observable profile behavior and keep deterministic construction.
• Add tests proving default bias is stationary and the stress profile drifts.
• Run formatting, checks, tests, and clippy applicable to the driver crate.

[tribbloid]

Issue #3 Review Verdict: Implementation Approved

Criteria Verification (all pass)

Criterion	Status	Evidence
DummyConfig::default has zero hard-iron drift	PASS	hard_iron_drift: ZERO at dummy.rs:90
Legacy amplitudes (2.0, 1.5, 2.5) and 5-min period retained in explicit stress profile	PASS	DummyConfig::adaptive_calibration_stress() at dummy.rs:105-111 uses ADAPTIVE_CALIBRATION_HARD_IRON_DRIFT and ADAPTIVE_CALIBRATION_DRIFT_PERIOD_US
Observable behavior and deterministic construction documented	PASS	Doc comments on hard_iron_drift, hard_iron_drift_period_us, �daptive_calibration_stress(); src/sim/AGENTS.md updated
Deterministic tests prove stationary default bias and stress-profile drift	PASS	default_hard_iron_bias_is_stationary and �daptive_calibration_stress_profile_drifts_deterministically in ests/dummy_profiles.rs
API is smallest coherent shape	PASS	One new public method, no new public types or fields

Validation Outcomes

Command	Result	Notes
cargo fmt --all -- --check	PASS (changed files)	Pre-existing diffs in examples/ and unrelated src/ files only; zero diffs in src/sim/dummy.rs, src/dummy_tests.rs, src/sim/AGENTS.md
cargo check --lib --all-features	PASS	43 warnings, all pre-existing in usion/, connection.rs, fi.rs, etc. Zero warnings in changed files
cargo test --test dummy_profiles --all-features	PASS	2/2 passed: default_hard_iron_bias_is_stationary, �daptive_calibration_stress_profile_drifts_deterministically
cargo clippy --lib --all-features -- -D warnings	FAIL (pre-existing)	65 errors, all pre-existing; zero warnings/errors from src/sim/dummy.rs or src/dummy_tests.rs

Unrelated Baseline Failures

• cargo check --all-targets --all-features: Fails due to pre-existing error in src/fusion/mag_calibration_test.rs referencing non-existent enum variant BadMagDataCause::DegenerateCalibrationSamples
• cargo clippy --lib --all-features -- -D warnings: 65 pre-existing errors across usion/, connection.rs, fi.rs, grawoow.rs,
  okid.rs (naming conventions, missing docs, dead code, etc.)
• cargo fmt --all -- --check: Pre-existing formatting diffs in examples/connection_sync.rs, examples/sensor_fusion.rs, src/connection.rs, src/ffi.rs, and newline style issues in multiple files

Additional Implementation Notes

The diff also resolves TODO.md high-severity issue #1 (non-planar trajectory) via the angular rate schedule mechanism, and medium-severity issue #2 (noise-balance) by raising default body rate to 20 RPM. Both are well-tested by gyro_matches_each_integrated_angular_rate and rajectory_covers_roll_pitch_yaw_and_non_planar_magnetometer_space.

No Issue #3 Defects Found

No recursive fix cycle needed.

---

GitHub action result: Attempted GraphQL deleteIssue on node I_kwDOLUBv2s8AAAABIaPGaQ (#3). Response: FORBIDDEN: Viewer not authorized to delete. Issue left open per policy.